Characterisation of nematode symbiotic bacteria and the in vitro liquid culture of Heterorhabditis zealandica and Steinernema yirgalemense

(1)

Heterorhabditis zealandica and Steinernema yirgalemense

Tiarin Ferreira

March 2013

Dissertation presented for the degree of

Doctor in Philosophy in the

Faculty

AgriSciences

at

Stellenbosch University

Supervisor: Dr Antoinette P Malan

(2)

Declaration

By submitting this thesis/dissertation electronically, I declare that the entirety of the work contained therein is my own, original work, that I am the sole author thereof (save to the extent explicitly otherwise stated), that reproduction and publication thereof by Stellenbosch University will not infringe any third party rights and that I have not previously in its entirety or in part submitted it for obtaining any qualification.

Date: March 2013

(3)

Abstract

Entomopathogenic nematodes have the potential to be outstanding biocontrol agents against agricultural pest insects. Combined with their bacterial symbionts, these biocontrol agents have proven to be very effective against numerous pests. The nematodes belong to the families Steinernematidae and Heterorhabditidae, and are ideal to be used in, and integrated with, pest management systems. There is a dire need for new and innovative methods to control agricultural pests, as numerous pest insects have developed resistance against broad-spectrum insecticides. Together with the environmental impact of these insecticides and the safety aspect regarding humans and animals, the need to develop new technologies, including entomopathogenic nematodes for pest management, is high. In this study, the associated symbiotic bacteria of three entomopathogenic nematodes species were isolated, and the potential of two nematode species to be successfully mass cultured in liquid medium was evaluated.

Regarding the symbiotic bacteria, results from the study showed that bacteria species from all three nematode species, Heterorhabditis noenieputensis, Steinernema khoisanae and Heterorhabditis zealandica, were novel. Heterorhabditis noenieputensis was isolated in the Mpumalanga province during a previous survey conducted in citrus orchards. The bacterium isolated from this nematode belongs to the genus Photorhabdus, and bear closest similarity (98.6%) to the type strain of P. luminescens subsp laumondii (TT01T). Photorhabdus luminescens subsp. noenieputensis subsp. nov., derives its name from the area where the nematode was sourced, namely the farm Springbokvlei, near the settlement Noenieput close to the Namibian border. Thus far, 85 Steinernema spp. have been described worldwide, including S. khoisanae which was isolated in the Western Cape province of South Africa. Four S. khoisanae strains, namely SF87, SF80, SF362 and 106-C, were used for characterisating the new bacteria from different localities in South Africa. Using the neighbor-joining method, all the strains were aligned with 97% homology to the 16S rRNA sequences of several Xenorhabdus- type strains, indicating that they belonged to the same genus. The multigene approach was used to distinguish between the Xenorhabdus spp. and partial recA, dnaN, gltX, gyrB and infB gene sequences of the various strains were analysed. The bacterium species was named Xenorhabdus khoisanae sp. nov. after the nematode from which it was isolated. The results showed that the third bacterium species, which was isolated from H.

(4)

zealandica, was new. The sequence of the bacteria strain clustered with the type strains of P. temperata and P. asymbiotica, indicate that it belonged to the genus Photorhabdus. This is the first study to show that H. zealandica associates with a luminescent Photorhabdus species, rather than with the known non-luminescent P. temperata.

The potential of H. zealandica and Steinernema yirgalemense mass culture in liquid was investigated. Results illustrated that H. zealandica and its P. luminescens symbiont can be successfully cultured in liquid. However, two generations occurred during the process time, instead of the desirable one-generation. The growth curve of the symbiotic bacteria during the process time was measured, in order to determine when the stationary phase was reached, with the results showing this to occur after 36 h. Therefore, the optimum amount of time required for inoculating the IJs and for aiding in maximum infective juvenile (IJ) recovery is 36 h for adding the nematodes post pre-culturing of the bacteria. Future research goals should be to increase the percentage recovery in liquid culture, which would increase the number of nematodes produced per ml, which would, therefore, reduce the processing time significantly.

The results from mass culturing the second nematode species, S. yirgalemense, indicated an asynchronous nematode development in the first generation. Growth curves were performed with the symbiotic bacteria that showed the exponential phase of Xenorhabdus started after 15 h, and that, after 42 h, the stationary phase was reached, with an average of 51 × 107 cfu·ml-1. Bioassays were performed to compare the virulence between in vitro- and in vivo-produced nematodes, with the results showing that the in vitro-produced nematodes were significantly less virulent than were the nematodes produced in vivo. The success obtained with the production of S. yirgalemense in liquid culture can serve as the first step in the optimising and upscaling of the commercial production of nematodes in industrial fermenters.

The last aim of the current study was to determine when Xenorhabdus reached the stationary phase, when it is grown in a 20-L fermenter, as this would be the optimum time at which to add the IJs of S. yirgalemense. Such characteristics as the effect of stationary phase conditions on the bacterial cell density and on the DO2 rate in the fermenter were investigated. The results showed that the stationary phase of Xenorhabdus was reached after 36 h at 30˚C, which took 6 h less than did the same procedures followed with the Xenorhabdus sp. cultured in Erlenmeyer flasks on orbital shakers. This is the first step

(5)

toward the liquid mass culturing of S. yirgalemense in industrial-size fermenters. Data from this study indicated the optimum amount of time that is required for adding nematodes to the bacterial culture in the fermenter, and for ensuring the optimum recovery of IJs, as well as a subsequent high yield of nematodes within a minimum processing time.

This is the first report of its kind to investigate comprehensively the successful liquid culture of two South African entomopathogenic nematode species for the sole purpose of evaluating potential commercialisation. Results emanating from this study could be used as groundwork in future, in combination with similar research such as culturing nematodes intensively in large fermenters.

(6)

Opsomming

Entomopatogeniese nematodes het die potensiaal om as doeltreffende biologiese beheeragente teen sleutelplaaginsekte gebruik te word. Elke nematood werk interaktief met ‘n spesifieke bakterium. Entomopatogeniese nematodes, behorende tot die families Steinernematidae en Heterorhabditidae, is ideale kandidate vir gebruik in ‘n geïntegreerde plaagbestuurprogram. Tans is daar ŉ behoefte vir nuwe metodes vir die beheer van plaaginsekte, omdat meeste insekte reeds weerstand opgebou het teen bestaande plaagdoders. As gevolg van die negatiewe impak van plaagdoders op die omgewing, asook kommer oor veiligheid vir die mens en diere, is die ontwikkeling en gebruik van alternatiewe plaagbeheermiddels noodsaaklik.

In die eerste deel van die studie word drie nuwe bakterie spesies geïsoleer en beskryf. Resultate van hierdie studie het aangetoon dat die bakterië spesies vanuit die nematode spesies, Heterorhabditis noenieputensis, Steinernema khoisanae, en Heterorhabditis zealandica, tot dusver onbeskryf was. Eersgenoemde, H. noenieputensis, is afkomstig van ŉ sitrusboord in die Mpumalanga Provinsie. Die bakterie hieruit geïsoleer behoort tot die genus Photorhabdus en is biologies verwant (98.6%) aan P. luminescens subsp laumondii (TT01T). Die bakterie is benaam as Photorhabdus luminescens subsp. noenieputensis nov. en is na die nematood waaruit dit geïsoleer is vernoem. Tot dusver is wêreldwyd 82 spesies van Steinernema spp. beskryf, insluitende S. khoisanae van die Weskaap provinsie. Vier bakterie isolate is van S. khoisanae, SF87, SF80, SF362 en 106-C geïsoleer. Die buur-koppeling metode was gebruik om te bepaal dat hierdie bakterie isolate tot 97% ooreenstem met verskeie isolate van Xenorhabdus se 16S rRNA DNS volgordebepalings. Om tussen Xenorhabdus spp. te onderskei is ŉ multi-geen benadering gebruik deur gedeeltelike recA, dnaN, gltX, gyrB en infB DNS basispaar volgordebepalings van die verskeie isolate te bepaal. Hierdie bakterie isolaat is soortgelyk ook vernoem as, Xenorhabdus khoisanae sp. nov., na die nematood waaruit dit geïsoleer is. Die derde onbekende bakteriële spesie is uit H. zealandica geïsoleer. Die DNS basispaar volgordebepaling van die 16S geen van SF41 toon aan dat dit in dieselfde groep as P. temperata en P. asymbiotica val en sodoende aan die genus Photorhabdus behoort. Hierdie is die eerste studie met die bevinding dat H. zealandica ook met ŉ ander bakterie spesie geassosieer kan word buiten die normale P. temperata spesie.

(7)

Die tweede deel van die studie gaan oor die teling van twee nematood spesies, H. zealandica en Steinernema yirgalemense, en hulle is geëvalueer vir hulle potensiaal om geteel te word in ŉ vloeibare medium. Die resultate het gewys dat H. zealandica met sy P. luminescens simbiont suksesvol in vloeistof aangeteel kan word, ten spyte van die feit dat daar twee generasies ontwikkel het, in plaas van die meer ideale enkel generasie. Die groeikurwe van die simbiotiese bakterie was gemonitor om te bepaal wanneer die stasionêre fase bereik word. Die resultate toon dat hierdie fase na 36 uur bereik was. Dus was die infektiewe nematode larwes eers na 36 uur tot die vloeibare medium waarin die bakterie geteel was bygevoeg. Navorsing in die toekoms moet dus gefokus wees om die persentasie herwinning van die infektiewe larwes te verhoog. Dit sal daartoe lei dat meer nematodes per ml geproduseer kan word en ook die prosesseringstyd van die nematodes verminder.

ŉ Tweede nematode spesie, S. yirgalemense, was ook in vloeistof geteel. Hier het ŉ asinkroniese ontwikkeling in die eerste generasie plaasgevind wat problematies is. Groeikurwes is bepaal van die bakteriële simbiont en die resultate het gewys dat die groeifase van Xenorhabdus na 15 uur in aanvang geneem het en dat die stasionêre fase bereik was na 42 uur met ŉ gemiddelde van 51 × 107

selle·ml-1. Die virulensie van nematodes wat in vitro geteel is, is vergelyk met die virulensie van nematodes wat in vivo geteel is en die resultate het getoon dat die in vitro geteelde nematodes minder virulent was. Die teling van S. yirgalemense in vloeistof was oor die algemeen meer suksesvol as die teling van H. zealandica in dieselfde medium.

Die doelwit van die laaste gedeelte van hierdie studie was om te bepaal wanneer Xenorhabdus die stasionêre fase bereik wanneer dit in ŉ 20-L fermenter gekweek word. Dit bepaal sodoende die optimale tyd wanneer die infektiewe larwes van S. yirgalemense bygevoeg behoort te word. Die uitwerking van die stasionêre fase op die bakteriële selle, asook die DO2-konsentrasie in die fermenter, was geëvalueer. Resultate het gewys dat die stasionêre fase van Xenorhabdus na 36 uur bereik was, wat 6 uur korter is as toe dit gekweek is in Erlenmeyer flesse. Hierdie studie is die eerste stap om die massa teling van S. yirgalemense in industriële fermenters suksesvol te bemeester. Die data wat verkry was, het aangedui wat die ideale tydsduur sal wees om die bakteriegetalle te vermeerder voordat die nematode bygevoeg word.

(8)

Hierdie is die eerste studie wat die teling van twee Suid-Afrikaanse nematode spesies omvattend in vloeistof evalueer het. Die hoof doelwit is om die potensiaal van hierdie nematode spesies, met die oog op kommersiële gebruik, te meet. Die resultate van hierdie studie kan gekombineer word met toekomstige studies in hierdie spesifieke navorsingsveld.

(9)

Acknowledgements

I wish to express my sincere appreciation to the following persons and institutions:

My promoters, Dr A. P. Malan, M. Addison and Dr. P. Addison for their guidance, interest and constructive criticism during the course of this study.

Prof. L. M. T. Dicks, Dr. C. van Reenen, Dr. A. Peters, Prof. R-U. Ehlers, Sylvie Pagès, P. Tailliez, A. Endo, C. Spröer, Prof. J. Gorgens, Dr. E.van Rensburg, E. Anane and Dr. S. Johnson for technical guidance and advice.

Prof. D. G. Nel for assistance with statistical analyses.

Prof. H. Geertsema for editing.

C. van Zyl, O.O. Kritzinger, C. Kapp, P. Le Vieux, Z. De Jager, S. Faure, Dr. J. Ross and E. Kritzinger for technical assistance and support.

The Department of Conservation Ecology and Entomology, Stellenbosch University.

The South African Apple and Pear Producer’s Association, Citrus Research International and The Technology and Human Resources Industry Programme for funding the project.

The National Research Foundation and German Academic Exchange Service for an additional bursary.

(10)

Table of Contents Abstract ... III Opsomming ... VI Acknowledgements ... IX CHAPTER 1 ... 1 Literature review ... 1

Xenorhabdus and Photorhabdus, bacterial symbionts of the entomopathogenic nematodes Steinernema and Heterorhabditis and their in vitro liquid culture ... 1

Introduction ... 1

Xenorhabdus and Photorhabdus ... 2

Life cycle of Xenorhabdus and Photorhabdus ... 3

Phenotypic variation ... 4

Taxonomy and systematics... 6

Heterorhabditis and Steinernema ... 9

Biology and behaviour of entomopathogenic nematodes ... 9

In vitro culturing of nematodes ... 11

Liquid culture process technology ... 12

Photorhabdus and Xenorhabdus have the ability to metabolise almost any kind of protein-rich medium. ... 12

Developmental biology of nematodes in liquid culture ... 14

Increasing recovery in liquid culture ... 17

References ... 18

CHAPTER 2 ... 34

Description of Photorhabdus luminescens subsp. noenieputensis subsp. nov., a symbiotic bacterium associated with a new Heterorhabditis species related to Heterorhabditis indica ... 34

Abstract ... 34

Materials and methods ... 36

Bacterial strains and growth conditions ... 36

Genotypic characterization... 36

Physiological and biochemical characterization ... 37

In vivo pathogenicity assay ... 38

Results... 38

(11)

In vivo pathogenicity assay ... 51

Discussion... 51

Acknowledgements ... 51

CHAPTER 3 ... 56

Description of Xenorhabdus khoisanae sp. nov., the symbiont of the entomopathogenic nematode Steinernema khoisanae ... 56

Abstract ... 56

Genotypic characterization... 59

Hybridization studies ... 60

Results... 60

Genotypic characterization... 64 Hybridization studies ... 71 Discussion... 72 Acknowledgements ... 72 References ... 72 CHAPTER 4 ... 76

Description of Photorhabdus sp. (SF41), a symbiont of the entomopathogenic nematode Heterorhabditis zealandica ... 76

Abstract ... 76

Material and methods ... 77

Physiological and biochemical characterisation ... 78

Genotypic characterisation... 78

Hybridisation studies ... 78

In vivo pathogenicity assay ... 79

Results... 79

(12)

Genotypic characterisation... 82

Hybridisation studies ... 87

In vivo pathogenicity assay ... 88

CHAPTER 5 ... 92

Development and population dynamics of Heterorhabditis zealandica and growth characteristics of the associated Photorhabdus luminescens symbiont in liquid culture ... 92

Abstract ... 92

Source of insects and nematodes ... 94

Isolation of symbiotic bacteria ... 94

Identification of symbiotic bacteria... 95

Axenisation of nematodes ... 95

Monoxenic culture protocol ... 96

Assessment of developmental stages ... 97

Bacterial growth curve ... 98

Virulence studies ... 98

Statistical analysis ... 99

Results... 99

Identity of symbiotic bacteria ... 99

Population dynamics of Heterorhabditis zealandica ... 99

Bacteria growth curve ... 102

Bioassays for virulence ... 104

CHAPTER 6 ... 114

Development and population dynamics of Steinernema yirgalemense and growth characteristics of the associated Xenorhabdus symbiont in liquid culture ... 114

Abstract ... 114

Source of insects and nematodes ... 117

(13)

Identification of associated symbiotic bacteria ... 118

Axenisation of nematodes ... 118

Monoxenic cultures ... 118

Assessment of developmental stages ... 119

Bacterial growth curve ... 119

Statistical analysis ... 120

Results... 121

Identification of the symbiotic bacteria ... 121

Population dynamics of S. yirgalemense ... 121

Bacteria growth curve ... 124

CHAPTER 7 ... 137

Investigating the growth characteristics of Xenorhabdus, associated with the entomopathogenic nematode Steinernema yirgalemense ... 137

Abstract ... 137

Isolation of symbiotic bacteria ... 140

Culture conditions ... 140

Assessment of bacterial growth ... 141

Statistical analysis ... 141 Results... 141 Discussion... 144 References ... 146 CHAPTER 8 ... 149 Conclusion ... 149

(14)

List of Tables

Table 2. 1. Main phenotypic characters differentiating Photorhabdus luminescens subspecies ... 49 Table 3.1. Carbohydrate reactions recorded for X. khoisanae and other Xenorhabdus species after 48 h of incubation at 28°C and using the BIOLOG GN microplate (Somvanshi et al., 2006; Tailliez et al., 2010). ... 62 Table 4.1. Results from API 20 NE and API 50 CH test strips for SF41, P. asymbiotica subsp. australis, P. asymbiotica subsp. asymbiotica and P. temperata subsp. temperata. ... 80 Table 4.2. Phenotypic characters differentiating SF41, P. asymbiotica subsp. australis, P. asymbiotica (Somvanshi et al., 2006). ... 82 Table 5.1. Body length for hermaphrodites, males, J1, J2, IJ, J3 and J4 (mean ± standard error and range) (one way ANOVA; F(5, 95 = 265.93; ρ = < 0,0001). ... 102 Table 6.1. Body length for females, males, IJ, J1, J2, J3 and J4 (mean ± standard error and range) (one-way ANOVA; F(5, 95 = 362.09; ρ = < 0,0001). ... 124 Table 7.1. Dissolved oxygen (DO2) and revolutions per minute (rpm) recording every 6 h during the growth phase of Xenorhabdus in a 20-L industrial fermenter. ... 144

(15)

Table of Figures

Fig. 2.1. Maximum likelihood (ML) phylogenetic tree of Photorhabdus luminescens calculated from five concatenated protein-coding sequences (recA, gyrB, dnaN, gltX and infB). Photorhabdus luminescens subsp. noenieputensis AM7T belongs to a monophyletic group including P. luminescens subsp. caribbeanensis, P. luminescens subsp. akhurstii and P. luminescens subsp. hainanensis. The ML analysis was carried out using the General Time Reversible model of substitution with gamma-distributed rate heterogeneity and a proportion of invariant sites determined for all five protein coding sequences determined by jModelTest to best fit with the data using the AIC criterion (Posada & Crandall, 1998). The concatenated sequences of Xenorhabdus bovienii strain T228T, Xenorhabdus nematophila strain ATCC19061T and Proteus mirabilis were used as outgroups. Bootstrap values (percentages of 100 replications) (Felsenstein, 1988) of more than 50% are shown at the nodes. The bar represents 5% divergence. ... 41 Fig. 2.2. Distance tree based on 16S rDNA sequences of Photorhabdus strains including P. luminescens subsp. noenieputensis subsp. nov. strain AM7T. The 16S rDNA sequences of P. luminescens subsp. africanis AM7T and P. luminescens subsp. laumondii TT01T share 98.6% nucleotide identity on a length of 1159 nucleotides. The neighbour-joining tree (Saitou & Nei, 1987) was constructed using the Kimura 2-parameter model (Kimura, 1980). Bootstrap values (Felsentein, 1988) of more than 50% are indicated at the nodes. Bar indicates 1 % sequence divergence. GenBank accession numbers of the sequences are in brackets. The sequences of X. nematophila ATCC19061T and P. luminescens subsp. laumondii TT01T were from http://www.cns.fr/agc/microscope/home/index.php. The sequence of P. mirabilis was from GenBank [NC010554]. ... 42 Fig. 2.3. ML tree based on recA sequences of Photorhabdus luminescens strains including P. luminescens subsp. noenieputensis AM7T. The ML analysis was carried out with the GTR model of substitution with gamma distributed rate heterogeneity and a proportion of invariant sites determined by jModelTest to best fit with the data using the AIC criterion (Posada & Crandall, 1998). The sequences of Xenorhabdus bovienii T228T, Xenorhabdus nematophila ATCC19061T and Proteus mirabilis were used as outgroups. Bootstrap values (Felsentein, 1988) of more than 50% are indicated at the nodes. Bar represents 5% divergence. GenBank accession numbers of the sequences are in brackets. The

(16)

sequences of X. nematophila ATCC19061T and P. luminescens subsp. laumondii TT01T were from http://www.cns.fr/agc/microscope/home/index.php. The sequence of P. mirabilis was from GenBank [NC010554]. ... 43 Fig. 2.4. ML tree based on gyrB sequences of Photorhabdus luminescens strains including P. luminescens subsp. noenieputensis AM7T. The ML analysis was carried out with the GTR model of substitution with gamma distributed rate heterogeneity and a proportion of invariant sites determined by jModelTest to best fit with the data using the AIC criterion (Posada & Crandall, 1998). The sequences of Xenorhabdus bovienii T228T, Xenorhabdus nematophila ATCC19061T and Proteus mirabilis were used as outgroups. Bootstrap values (Felsentein, 1988) of more than 50% are indicated at the nodes. Bar represents 5% divergence. GenBank accession numbers of the sequences are in brackets. The sequences of X. nematophila ATCC19061T and P. luminescens subsp. laumondii TT01T were from http://www.cns.fr/agc/microscope/home/index.php. The sequence of P. mirabilis was from GenBank [NC010554]. The sequence of P. luminescens subsp. kleinii strain KMD37 (= DSM23513) is that published by An & Grewal (2011). ... 44 Fig. 2.5. ML tree based on dnaN sequences of Photorhabdus luminescens strains including P. luminescens subsp. noenieputensis AM7T. The ML analysis was carried out with the GTR model of substitution with gamma distributed rate heterogeneity and a proportion of invariant sites determined by jModelTest to best fit with the data using the AIC criterion (Posada & Crandall, 1998). The sequences of Xenorhabdus bovienii T228T, Xenorhabdus nematophila ATCC19061T and Proteus mirabilis were used as outgroups. Bootstrap values (Felsentein, 1988) of more than 50% are indicated at the nodes. Bar represents 10% divergence. GenBank accession numbers of the sequences are in brackets. The sequences of X. nematophila ATCC19061T and P. luminescens subsp. laumondii TT01T were from http://www.cns.fr/agc/microscope/home/index.php. The sequence of P. mirabilis was from GenBank [NC010554]. ... 45 Fig. 2.6. ML tree based on gltX sequences of Photorhabdus luminescens strains including P. luminescens subsp. noenieputensis AM7T. The ML analysis was carried out with the GTR model of substitution with gamma distributed rate heterogeneity and a proportion of invariant sites determined by jModelTest to best fit with the data using the AIC criterion (Posada & Crandall, 1998). The sequences of

(17)

Xenorhabdus bovienii T228T, Xenorhabdus nematophila ATCC19061T and Proteus mirabilis were used as outgroups. Bootstrap values (Felsentein, 1988) of more than 50% are indicated at the nodes. Bar represents 5% divergence. GenBank accession numbers of the sequences are in brackets. The sequences of X. nematophila ATCC19061T and P. luminescens subsp. laumondii TT01T were from http://www.cns.fr/agc/microscope/home/index.php. The sequence of P. mirabilis was from GenBank [NC010554]. ... 46 Fig. 2.7. ML tree based on infB sequences of Photorhabdus luminescens strains including P. luminescens subsp. noenieputensis AM7T. The ML analysis was carried out with the GTR model of substitution with gamma distributed rate heterogeneity and a proportion of invariant sites determined by jModelTest to best fit with the data using the AIC criterion (Posada & Crandall, 1998). The sequences of Xenorhabdus bovienii T228T, Xenorhabdus nematophila ATCC19061T and Proteus mirabilis were used as outgroups. Bootstrap values (Felsentein, 1988) of more than 50% are indicated at the nodes. Bar represents 5% divergence. GenBank accession numbers of the sequences are in brackets. The sequences of X. nematophila ATCC19061T and P. luminescens subsp. laumondii TT01T were from http://www.cns.fr/agc/microscope/home/index.php. The sequence of P. mirabilis was from GenBank [NC010554]. ... 47 Fig. 3.1. Bootstrap percentages above 55% are given at branching points. Phylogenetic relationship of strain Steinernema khoisanae sp. nov. SF87T to known Xenorhabdus spp. based on 16S rRNA gene sequences. The tree was constructed by the neighbour-joining method. Photorhabdus asymbiotica subsp. asymbiotica was used as an outgroup. Bootstrap percentages above 70% are given at branching points. ... 65 Fig. 3.2. Phylogenetic relationship of strains SF87T, SF80, 106-C and SF362 to known Xenorhabdus spp. based on recA gene sequences. The tree was constructed by the neighbour-joining method. Photorhabdus asymbiotica subsp. asymbiotica was used as an outgroup. Bootstrap percentages above 70% are given at branching points. ... 66 Fig. 3.3. Phylogenetic relationship of strains SF87T, SF80, 106-C and SF362 to known Xenorhabdus spp. based on dnaN gene sequences. The tree was constructed by the neighbour-joining method.

(18)

Photorhabdus asymbiotica subsp. asymbiotica was used as an outgroup. Bootstrap percentages above 70% are given at branching points. ... 67 Fig. 3.4. Phylogenetic relationship of strains SF87T, SF80, 106-C and SF362 to known Xenorhabdus spp. based on gltX gene sequences. The tree was constructed by the neighbour-joining method. Photorhabdus asymbiotica subsp. asymbiotica was used as an outgroup. Bootstrap percentages above 70% are given at branching points. ... 68 Fig. 3.5. Phylogenetic relationship of strains SF87T, SF80, 106-C and SF362 to known Xenorhabdus spp. based on gyrB gene sequences. The tree was constructed by the neighbour-joining method. Photorhabdus asymbiotica subsp. asymbiotica was used as an outgroup. Bootstrap percentages above 70% are given at branching points. ... 69 Fig. 3.6. Phylogenetic relationship of strains SF87T, SF80, 106-C and SF362 to known Xenorhabdus spp. based on infB gene sequences. The tree was constructed by the neighbour-joining method. Photorhabdus asymbiotica subsp. asymbiotica was used as an outgroup. Bootstrap percentages above 70% are given at branching points. ... 70 Fig. 3.7. Phylogenetic relationship of strains SF87T, SF80, 106-C and SF362 to known Xenorhabdus spp. based on concatenated (16S rRNA, recA, dnaN, gltX, gyrB and infB) gene sequences. The tree was constructed by the neighbour-joining method. Photorhabdus asymbiotica subsp. asymbiotica was used as an outgroup. Bootstrap percentages above 70% are given at branching points. ... 71 Fig. 4.1. Phylogenetic tree based on 16S rDNA sequences of Photorhabdus strains, including strain SF41. The tree was constructed using the maximum-likelihood method. Escherichia coli was used as an outgroup. ... 83 Fig. 4.2. Phylogenetic tree based on recA sequences of Photorhabdus strains, including strain SF41. The tree was constructed using the maximum-likelihood method. Xenorhabdus hominickii was used as an outgroup. ... 84 Fig. 4.3. Phylogenetic tree based on gyrB sequences of Photorhabdus strains, including strain SF41. The tree was constructed using the maximum-likelihood method. Xenorhabdus hominickii was used as an outgroup. ... 85

(19)

Fig. 4.4. Phylogenetic tree based on dnaN sequences of Photorhabdus strains, including strain SF41. The tree was constructed using the maximum-likelihood method. Xenorhabdus hominickii was used as an outgroup. ... 86 Fig. 4.5. Phylogenetic tree based on gltX sequences of Photorhabdus strains, including strain SF41. The tree was constructed using the maximum-likelihood method. Xenorhabdus hominickii was used as an outgroup. ... 87 Fig. 5.1. Mean population density (95% confidence interval) of males, females and hermaphrodites of Heterorhabditis zealandica in monoxenic liquid culture at 25˚C, over a period of 16 days (one-way ANOVA; F(30, 405) = 61.81)... 100 Fig. 5.2. Mean population density (95% confidence interval) of J1/J2 and J3/J4 larvae of Heterorhabditis zealandica in monoxenic liquid culture at 25˚C, over a period of 16 days (one-way ANOVA; F(15, 270) = 31.12). ... 101 Fig. 5.3. Mean population density (95% confidence interval) of Heterorhabditis zealandica IJs in monoxenic liquid culture at 25˚C, over a period of 16 days (one-way ANOVA; F(15, 135) = 4133)... 101 Fig. 5.4. The mean colony forming units per ml (95% confidence interval) of Photorhabdus luminescens in trypticase soy broth, over a period of 48 h, at 30˚C in the dark (one-way ANOVA; F(16, 112) = 168.4). . 103 Fig. 5.5. The mean optical density (measured at 595 nm) (95% confidence interval) of Photorhabdus luminescens in trypticase soy broth, over a period of 48 h at 30˚C in the dark (one-way ANOVA; F(16, 32) = 132.5). ... 103 Fig. 5.6. The mean colony forming units per ml(95% confidence interval) of Photorhabdus luminescens in liquid culture over a period of 16 days, post IJ inoculation at 25˚C (one-way ANOVA; F(15, 105) = 184.2). 104 Fig. 5.7. The mean percentage mortality (95% confidence interval) of Galleria mellonella larvae inoculated with 200 IJs/insect of Heterorhabditis zealandica after two days (one-way ANOVA; F(1, 68) = 28.26; ρ < 0,0001). ... 105 Fig. 6.1. Mean population density (95% confidence interval) of Steinernema yirgalemense in monoxenic liquid culture at 25˚C, over a period of 15 days. The density of males, females, pre-adult stage juveniles (J3 and J4 stages) and endotokia matricida females is indicated (one-way ANOVA; F(42, 504) = 52.12). . 122

(20)

Fig. 6.2. Mean population density (95% confidence interval) of J1 and J2 stages of Steinernema yirgalemense, in monoxenic liquid culture at 25˚C, over a period of 15 days (one-way ANOVA; F(14, 252) = 26.89). ... 123 Fig. 6.3. Mean population density (95% confidence interval) of infective juveniles (IJ) and pre-infective juveniles (J2d) of Steinernema yirgalemense, in monoxenic liquid culture at 25˚C, over a period of 15 days (one-way ANOVA; F(14, 126) = 411.12). ... 123 Fig. 6.4. The colony forming units·ml-1 (95% confidence interval) of Xenorhabdus sp. in trypticase soy broth, over a period of 48 h, at 30˚C in the dark (one-way ANOVA; F(16, 112) = 206.7). ... 125 Fig. 6.5. The optical density (measured at 595 nm) (95% confidence interval) of Xenorhabdus sp. in Luria broth, over a period of 48 h, at 30˚C in the dark (one-way ANOVA; F(16, 32) = 248.1). ... 125 Fig. 6.6. The colony forming units (95% confidence interval) ml-1 of Xenorhabdus sp. in liquid culture over a period of 15 days, at 25˚C (one-way ANOVA; F(14, 98) = 29.7). ... 126 Fig. 6.7. The mean percentage mortality (95% confidence interval) of Galleria mellonella larvae after .. 127 two days of inoculation with 200 IJs/insect of Steinernema yirgalemense (one-way ANOVA: F(1,68) = 41.783; ρ = < 0.0001). ... 127 Fig. 7.1. The cell density·ml-1 of Xenorhabdus spp. in liquid medium, over a period of 60 h in a 20-L fermenter at 30˚C. ... 142 Fig. 7.2. The dissolved oxygen readings taken every 3 h, over a period of 60 h, in a 20-L fermenter, in which Xenorhabdus was cultured in the liquid medium at 30˚C ... 143 Fig. 7.3. Bacterial cell density and dissolved oxygen over a period of 60 h in a 20-L fermenter in which Xenorhabdus spp. was cultured in the liquid medium at 30˚C. ... 144

(21)

CHAPTER 1 Literature review

Xenorhabdus and Photorhabdus, bacterial symbionts of the entomopathogenic nematodes Steinernema and Heterorhabditis and their in vitro liquid culture

Introduction

Annually, numerous insect pests cause damage to fruit and vegetables that are grown as food crops all over the world (Wyniger, 1962). These insect pests are a serious economic burden on agriculture in South Africa. Control methods that are highly specific to the target pests and that are, in addition, environmentally friendly, such as biological control agents, should constitute a major component of integrated pest management systems.

Entomopathogenic nematodes (EPNs) represent an important part of the spectrum of potential biological control agents. Previous research in South Africa has shown that two local nematodes, Heterorhabditis zealandica Poinar, 1990 and Steinernema yirgalemense Nguyen, Tesfamariam, Gozel, Gaugler & Adams, 2005, in particular, have great insecticidal potential (De Waal et al., 2011a,b, 2012; Malan et al., 2011; Van Niekerk & Malan, 2012). Therefore, the ability to mass culture these two nematode species in liquid medium, using in vitro technology, is an important step toward their application as biocontrol agents on a commercial scale against key insect pests. However, for in vitro technology to be successful, the nematode-bacteria interaction needs to be understood. Bacterial symbionts of EPNs need to be isolated and characterised, with the life cycle of the nematode in culture requiring to be understood to be able to optimise for maximum nematode yield in liquid culture.

(22)

Xenorhabdus and Photorhabdus

EPNs of the families Steinernematidae and Heterorhabditidae share a mutualistic relationship with bacteria of the genera Xenorhabdus Thomas & Poinar, 1979 and Photorhabdus Boemare, Akhurst & Mourant, 1993, respectively. The bacteria belong to the family Enterobacteriaceae, whose characteristics include being gram-negative and having non-fermentative rods (Koppenhöfer, 2007). Photorhabdus and Xenorhabdus are a unique group, as they are phenotypically (Holt et al., 1994) and genotypically (Brenner & Farmer, 2005) similar to no other genera grouped in this family. Both of them produce the enterobacterial common antigen that is present among the species of Enterobacteriaceae (Ramia et al., 1982). Said nematode bacterial symbionts are pathogenic to insects. There is, however, an exception to the rule, as one species, Photorhabdus asymbiotica Fischer-Le Saux, Viallard, Brunel, Normund & Boemare, 1999 has been found to be an opportunistic pathogen to humans (Farmer et al., 1989; Peel et al., 1999).

The bacterial symbionts are carried by their associated nematodes and released into the haemolymph of a host insect. Once inside the insect, the symbiotic bacteria overcome the immune system of the host and release endo- and exotoxins. Septicaemia develops, with the death of the host insect usually occurring within one to two days (Poinar, 1990a; Forst & Clarke, 2002). The bacterial symbionts contribute to the symbiotic relationship with EPNs by creating conducive conditions for nematode growth and reproduction in the host (Boemare et al., 1997b). Nutrients, as well as antimicrobial substances, are provided that inhibit the growth of a wide range of micro-organisms (Akhurst, 1982) in and on the insect cadaver. They also excrete substances preventing scavenging nematodes and insects from utilising the cadaver as a food source (Zhou et al., 2002). Up to three generations of nematodes can be produced per host, depending on the size of the insect host (Kakouli-Duarte & Hague, 1999). In smaller hosts, only one or two generations are produced (Ferreira, 2010; Van Niekerk & Malan, 2012). As soon as the food in the host cadaver is depleted, a new cohort of infective juveniles (IJs) enters the environment, with their bacterial symbionts enclosed in the digestive system of the nematode. The bacteria have not so far been reported as occurring freely in nature, but only in association with the nematode (Akhurst et al., 2004; Hazir et al., 2004; Lengyel et al., 2005; Tailliez et al., 2006).

(23)

Previous research has been aimed at screening and assaying the insecticidal properties of several of the symbiotic bacteria. Currently, there are three described species of Photorhabdus and 15 subspecies and 23 species of Xenorhabdus. Since 2004, three new subspecies of Photorhabdus (Akhurst et al., 2004; Hazir et al., 2004) and 14 new species of Xenorhabdus (Lengyel et al., 2005; Tailliez et al., 2006) have been described. However, many bacterial isolates from previously described nematode species still require characterisation.

Life cycle of Xenorhabdus and Photorhabdus

The Heterorhabditidae and Steinernematidae families of nematodes are obligate insect pathogens. The only way in which such nematodes can persist outside the insect host is as specialised third-stage IJs. Their bacterial symbionts are contained in the intestinal tract of the IJs. The bacterial symbiont of Heterorhabditis, which is Photorhabus, mainly colonises the anterior region of the intestine just posterior to the basal bulb. However, it is also to varying degrees located throughout the remainder of the intestine (Endo & Nickle, 1991; Ciche & Ensign, 2003). Xenorhabdus is the bacterial symbiont of Steinernema, with the nematodes having a specialised bilobed intestinal vesicle that becomes colonised by the bacteria (Bird & Akhurst, 1983; Martens et al., 2003). In both of the nematode symbionts, the bacteria are in a dormant state inside the IJ, apart from in the case of Steinernema carpocapsae (Weiser, 1955) Wouts, Mráček, Gerdin & Bedding, 1982, where limited bacterial growth takes place until the intestinal vesicle is colonised by Xenorhabdus nematophila (Poinar & Thomas, 1965; Thomas & Poinar, 1979; Martens et al., 2003).

Once the IJ enters the haemocoel of a susceptible insect host, the nematode resumes development and releases its bacterial symbiont. Inside the insect gut, X. nematophila cells are released from the vesicle into the nematode’s intestine (Sicard et al., 2004) by means of defecation (Poinar & Thomas, 1966; Wouts, 1984; Martens et al., 2004; Sicard et al., 2004). Photorhabdus are released through the mouth of the nematode, in an action resembling regurgitation (Ciche & Ensign, 2003).

The nematodes and bacteria work together to overcome the immune response of the host, thus allowing the bacteria to proliferate (Koppenhöfer, 2007). Steinernema carpocapsae secretes proteins suppressing the immune response of the insect host and this may aid the release of their symbionts (Gotz

(24)

et al., 1981; Simoes, 1998). It is unknown whether similar proteins are released by Heterorhabditis (Forst & Clarke, 2002). Both genera produce haemolysin activity (Brillard et al., 2001, 2002). While the bacteria develop in the insect host, they produce toxins and exo-enzymes. This results in the septicaemia of the insect host and the cadaver, which provides nutrition for the developing nematodes (Forst & Clarke, 2002). The above takes place early on in the infection and preceding the insect’s death (Sicard et al., 2004). Likewise, Photorhabdus reproduce in the haemocoel of Manduca sexta Linnaeus, 1763, destroying the immune system. Toxins are released by the bacteria late in the infection, destroying the epithelium of the midgut (Bowen et al., 1998; Silva et al., 2002).

Towards the end of bacterial growth, the symbionts produce a range of antimicrobial compounds that protect the cadaver from colonisation by other organisms. The compounds include bacteriocins, which are active against closely related bacteria. In P. luminescens, the bacteriocins are also active against distantly related bacterial taxa (Thaler et al., 1995; Sharma et al., 2002). Antibiotics are other compounds that are produced by the bacteria, which are active against fungi, yeasts and other bacteria (Akhurst, 1982; Boemare et al., 1997a; Webster et al., 2002).

Developing nematodes feed on a mixture of bacteria and bioconverted host tissue, enabling them to produce one to three generations until the food resources in the cadaver are depleted. As soon as depletion takes place, the nematodes develop a new generation of a special third generation of IJs enclosing the symbiotic bacterial cells, which exit the cadaver in search of a new susceptible insect host (Koppenhöfer, 2007).

Phenotypic variation

Phenotypic variants are produced by both Xenorhabdus and Photorhabdus. The primary form, which is called form I, is associated with the nematodes. The secondary form of the bacteria (Form II cells) arises abruptly when in artificial culture, and seldom occurs in the insect host during the later stages of nematode reproduction (Akhurst, 1980). The two forms of bacteria mentioned differ both morphologically and physiologically.

(25)

Form I cells are larger, as well as motile, as the result of peritrichous flagella, which form II cells do not have (Givaudan et al., 1995). Form I cells are able to absorb certain dyes and to produce crystalline inclusion bodies, antibiotics, lipase, and protease, while certain strains of Photorhabdus are also bioluminescent (Akhurst, 1980, 1982; Couche et al., 1987; Boemare & Akhurst, 1988; Forst et al., 1997). All of these mentioned characteristics are reduced or missing in the variant cells. Form II cells in X. nematophila do not produce a stationary-phase outer membrane protein called OpnB (Volgyi et al., 2000). Xenorhabdus form II colonies have little or no pigment on nutrient agar, while Photorhabdus form II colonies depend on the strain or species, which differ (Akhurst, 1983; Akhurst & Boemare, 1988; Boemare & Akhurst, 1988; Boemare et al., 1997a).

Higher levels of respiratory enzyme activity are present in form II cells for both X. nematophila and P. luminescens, and such cells are also more capable of taking up nutrients than are other cells (Smigielski et al., 1994). Differences in pathogenicity exist between the phenotypic variants of X. nematophila in lepidopteran hosts, but the overall pathogenicity is maintained in form II cells. Form I and form II cells are both pathogenic against Galleria mellonella (L.) (Lepidoptera: Pyralidae) (Akhurst, 1980). Photorhabdus form II cells cannot support the growth and reproduction of Heterorhabditis (Gerritsen & Smits, 1993, 1997). On the contrary, form II cells of X. nematophila can support nematode reproduction both in vitro (Ehlers et al., 1990; Volgyi et al., 2000) and in vivo (Sicard et al., 2005). In S. carpocapsae, the production of xenorhabdicin is maintained in form II cells, but, when antagonistic bacteria are present, they are not sensitive to xenorhabdicin, with form II cells providing less protection than do form I cells to their nematode host (Sicard et al., 2005). Producing antibiotics is costly with regard to metabolism, and, without this function, a great deal more nutrient uptake can occur than with it. The above ensues with increased proliferation of the bacteria and, as a result, they become more adaptive for survival (Smigielski et al., 1994; Sicard et al., 2005). Phenotypic change and the mechanisms that drive it are in the process of becoming understood. For example, form I characteristics in P. luminescens are negatively regulated by HexA (Joyce & Clarke, 2003), with, in X. Nematophila, such characteristics being positively regulated by Lrp (Park et al., 2007). Mutualism and pathogenesis are both affected by said bacterial regulators, and will occasionally revert to form I bacteria in Xenorhabdus, which has, however, not yet been documented for Photorhabdus (Givaudan et al., 1995; Forst & Clarke, 2002).

(26)

Taxonomy and systematics

According to Stackebrandt et al. (2002), the current definition of bacterial species, although arbitrary and artificial, is still universally operational. A genomically consistent group of individual isolates sharing a high degree of similarity is regarded as a species. The degree of similarity should be present in numerous independent features, as well as be diagnosable by means of the presence of a discriminative phenotype (Rossello-Mora & Amann, 2001; Stackebrandt et al., 2002; Adams et al., 2006). The main criterion for the description of bacterial species continues to be DNA : DNA homology. Strains within a species should preferably have a DNA : DNA relatedness value of 70% or higher and a ΔTm of 5˚C or lower (Wayne et al., 1987; Rossello-Mora & Amann, 2001; Stackebrandt et al., 2002). Such values, however, are not absolute for the description of a new species (Rossello-Mora & Amann, 2001), as additional molecular techniques are encouraged when the degree of similarity of DNA : DNA reassociation is adequate (Stackebrandt et al., 2002). As a result of the low DNA : DNA relatedness values of earlier studies, and the differences between 16S rRNA gene sequences (Boemare et al., 1993; Nishimura et al., 1994), previous Xenorhabdus species have been described without considering DNA : DNA reassociation (Lengyel et al., 2005; Somvanshi et al., 2006). According to Tailliez et al. (2006), a consistent alternative to DNA : DNA hybridisation is a combination of randomly amplified polymorphic DNA (RAPD) and enterobacterial repetitive intergenic consensus (ERIC) sequences.

Multi-locus sequence typing (MLST) is a method that holds great promise for the delineation of species. Partial sequences of internal fragments from multiple housekeeping genes are used. Based on the number of different loci, the evolutionary distance between isolates is subsequently quantified (Maiden et al., 1998; Adams et al., 2006). The method concerned can easily be replicated and, furthermore, there are publicly accessible databases containing sequences and software to use for comparing specific isolates (see http://www.mlst.net and http://pubmlst.org). The genomic association of strains can be determined with more confidence using MLSTs than with the use of DNA : DNA reassociation (Lan & Reeves, 2001; Adams et al., 2006). When using MLSTs the concatenation of several of these unlinked gene sequences have the ability to yield more robust phylogenetic trees (Rokas et al., 2003; Wertz et al., 2003) when compared to single-gene phylogenies. According to Tailliez et al (2010), creating a resolved phylogeny of these bacteria is necessary in order to study their co-evolution

(27)

with their nematode hosts. Analysis of the topology of single-gene phylogenetic trees (Doolittle, 1999) can be used to identify plausible lateral gene transfers (LGT) between species, which can have implications for the classification of strains and new isolates in the genera Photorhabdus and Xenorhabdus (Tailliez et al., 2010).

A preferred method to DNA similarity or molecular methodology is a polyphasic approach for species description. The description of species should include an almost complete 16S rDNA sequence, the G + C mol% content of the type strain of the type genus, the phenotype and also the chemotaxonomic characters (Stackebrandt et al., 2002). Standard systems can easily be used to obtain phenotypic data. The description of Xenorhabdus and Photorhabdus species has been done in toto using API® substrate panels, which were designed to illustrate the carbohydrate metabolism. Recently, Biolog GN™ has been used for species description and bacterial identification, in order to illustrate substrate utilisation (Hazir et al., 2004; Lengyel et al., 2005; Gouge & Snyder, 2006; Somvanshi et al., 2006). As phenotypes that are described by metabolism only are regarded as insufficient, supplementary chemotaxonomic characters should be considered (Rossello-Mora & Amann, 2001).

The considerable differences in biochemical reactions for Xenorhabdus and Photorhabdus that have been reported (Holt et al., 1994; Brenner & Farmer, 2005) have complicated the comparing of species, with the variation probably being a result of using different bacterial strains and/or phenotypic variants (Akhurst & Boemare, 1988). Other possibilities for the variations may include weak and slow reactions, and the media type that is used for biochemical characterisation (Holt et al., 1994).

The first bacterial symbiont that was isolated from the DD-136 strain of S. carpocapsae was described as a new bacterium species, namely Achromobacter nematophila Poinar & Thomas, 1965, by Poinar & Thomas (1965). The genus Achromobacter was later rejected (Hendrie et al., 1974) and reassigned to a different genus. The authors decided to create a new genus, Xenorhabdus, as A. nematophila did not closely enough resemble any of the accepted genera. The new genera included X. nematophilus, which is a symbiont of a Steinernema species, and X. luminescens, which is a symbiont of a Heterorhabditis species (Thomas & Poinar, 1979). A distinct difference could be discerned between X. luminescens and other Xenorhabdus strains, both in terms of their phenotypic (Akhurst, 1983; Akhurst &

(28)

Boemare, 1988; Boemare & Akhurst, 1988) and their molecular characteristics (Grimont et al., 1984; Farmer et al., 1989; Suzuki et al., 1990). Grimont et al. (1984) placed X. luminescens into a different group through the use of s1 nuclease and hydroxyapatite methods for determining DNA : DNA similarity. Photorhabdus was proposed as a new genus, as a result of insufficient DNA homology to other Xenorhabdus species (Boemare et al., 1993). Even though Photorhabdus is quite species poor in comparison to Xenorhabdus, it is still more homogenous than is the Xenorhabdus group (Akhurst et al., 1996). Most of the bacterial symbionts of recently isolated EPNs must still be described.

Xenorhabdus and Photorhabdus belong to the family Enterobacteriaceae (Rahn, 1937) Ewing, Farmer & Brenner, 1980, as well as the gamma subdivision of the Proteobacteria. Characteristics that said bacteria have include: the possession of gram-negative rods; motility by means of peritrichous flagella or non-motility; facultative anaerobism; negativity for oxidase; and asporogenous, non-acid fast, chemoorganic heterotrophs, with respiratory and fermentative metabolisms (Brenner, 1999; Brenner & Farmer, 2005). When considering the phenotypical characteristics of Xenorhabdus and Photorhabdus, they can be seen to be out of character, compared to other members of the Enterobacteriaceae family (Holt et al., 1994).

When comparing Xenorhabdus and Photorhabdus, two main differences are that the latter are catalase-positive, with the majority being bioluminescent. Xenorhabdus isolates are negative for both characteristics concerned (Poinar et al., 1980; Farmer, 1984; Boemare & Akhurst, 1988). Distinguishing the two groups of bacteria clearly is the fact that Photorhabdus contains unique sequences in the 16S small subunit rDNA, which Xenorhabdus lacks. The sequence TTCG of Xenorhabdus is at the 208–211 position (in terms of E. coli numbering), while Photorhabdus contains the longer TGAAG sequence instead (Szallas et al., 1997).

By assessing bacterial diversity and identification through the polymorphism of the gene coding for the ribosomal RNA subunit, laborious phenotypic characterisation can be avoided. A method, such as the restriction fragment analysis of PCR-amplified gene products, has been used successfully for such purposes. The identification of Xenorhabdus and Photorhabdus can accurately be identified by means of restriction fragment length polymorphisms (RFLP) of the 16S rRNA gene sequence (Brunel et al., 1997;

(29)

Fischer-Le Saux et al., 1998; Bonifassi et al., 1999). Distinguishing between Xenorhabdus and Photorhabdus is also practical when using 16S rRNA sequences (Liu et al., 2001; Sergeant et al., 2006).

Heterorhabditis and Steinernema

More than 30 nematode families are known to parasitise, or are associated with, insects (Nickle, 1972; Maggenti, 1981; Poinar, 1983, 1990b; Kaya & Stock, 1997). As a result of biocontrol potential, more research has been undertaken into seven families, including Mermithidae, Allantonematidae, Neotylenchidae, Sphaerularidae, Rhabditidae, Steinernematidae and Heterorhabditidae. The last two families have received the most attention so far, as they can be cultured and formulated, and can be used to control a wide range of insect pests within a short space of time (Lacey et al., 2001).

Phasmarhabditis hermaphrodita (Schneider) is a member of the family Rhabditidae. It is known to suppress numerous slug and snail species, and has been developed as a biological molluscicide (Wilson et al., 1993; Glen & Wilson, 1997; Wilson & Gaugler, 2000). Such potential biocontrol agents of plant-parasitic nematodes and plant pathogens as predatory mononchids, dorylaimids, nygolaimids, diplogasterids and the fungal-feeding nematode, Aphelenchus avenae Bastian, have also been studied (Kasab & Abel-Kader, 1996; Lootsma & Scholte, 1997; Matsunaga et al., 1997; Choudhury & Sivakumar, 2000), without much success.

Biology and behaviour of entomopathogenic nematodes

EPNs of the family Steinernematidae and Heterorhabditidae are lethal pathogens of insects. In nature, they play a role in regulating the natural population of insects, but their main point of interest is their inundative application as biocontrol agents. The unique partnership between the nematode and the lethal insect-pathogenic bacterium has helped to ensure their success as biocontrol agents (Griffin et al., 2005).

Even though Heterorhabditis and Steinernema have adopted the same lifestyle, they belong to different families (Heterorhabditidae and Steinernematidae respectively) (Blaxter et al., 1998). Similarities include their association with insect-pathogenic bacteria, in addition to which they are thought to have originated through convergent evolution (Poinar, 1993). Both Steinernema and Heterorhabditis have a

(30)

single free-living stage, the IJ stage, which carries bacteria from the genus Xenorhabdus and Photorhabdus, respectively, in its gut (Boemare et al., 1993). The IJ can enter the insect through its mouth, anus or spiracles and move towards the haemocoel. Some species have the ability to penetrate through the insect cuticle (Bedding & Molyneux, 1982; Peters & Ehlers, 1994). For example, Heterorhabditis are able to do so by means of their anterior dorsal tooth (Bedding & Molyneux, 1982).

Once the IJ is inside the haemocoel of the host, the nematode releases the cells of its bacterial symbiont from its intestine. The insect’s haemolymph is extremely nutrient-rich, and the bacteria multiply exponentially, causing insect death within 24-48 h. IJs feed on the proliferating bacteria and digested host tissue after they recover from their arrested state. Nematodes develop to the fourth larval stage, and subsequently to adult stages in order to reproduce. Depending on the available nutrients and resources, more than one generation can occur (Dix et al., 1992).

Steinernematids and heterorhabditids have a different mode of reproduction. The first generation for the latter consists of self-fertilising hermaphrodites, with males and females developing in subsequent generations (Dix et al., 1992). For steinernematids, all the generations can reproduce through amphimixis (i.e. cross-fertilisation involving males and females) (Poinar, 1990a). Recently, however, a Steinernema sp. was found, of which most of the individuals were self-fertilising hermaphrodites, with a small portion of the population in each generation being males (Griffin et al., 2001). Therefore, in general, when only a single IJ invades a host insect, a heterorhabditid is able to reproduce and develop, while most steinernematids require two individuals, one male and one female, to be present before they can reproduce (Griffin et al., 2005).

At first, eggs are laid by the females or hermaphrodites, and, in older females or hermaphrodites, the eggs hatch in the uterus, with the parental tissue being consumed by the juvenile. The process concerned is known as ‘endotokia matricida’ (Johnigk & Ehlers, 1999). Parental tissue makes for an extremely efficient conversion from insect biomass to IJ biomass. When there is an adequate food supply, the juveniles develop into adults. When the conditions inside the host insect are crowded, and there are limited resources, the IJs do not develop further. In a large insect host, hundreds of thousands of IJs can

(31)

develop, emerging from the insect cadaver over a period of days, whereupon they then begin to search for a new insect host (Griffin et al., 2005).

When an IJ has newly emerged from its insect host, it retains the second-stage cuticle as a sheath. The sheath assists with the prevention of desiccation, freezing and fungal pathogen infection, especially in heterorhabditids (Timper & Kaya, 1989; Campbell & Gaugler, 1991b; Wharton & Surrey, 1994). Steinernematids lose their sheath quite easily as they move through the soil, while the heterorhabditid sheath is not so easily lost, due to it being more tight-fitting (Campbell & Gaugler, 1991a; Dempsey & Griffin, 2003).

In vitro culturing of nematodes

The production of nematodes in vitro requires a detailed understanding of the biology and behaviour of the nematode species being mass produced. The first axenically liquid culture was concocted by Stoll (1952), using raw liver extract shaken in flasks. The use of bioreactors to culture nematodes was first attempted and described in 1986 by Pace et al. When said researchers cultured the nematodes in a standard 10-l bioreactor (Braun Biostat E), the main finding was that an impeller tip velocity of 1 m/s or more led to the disruption of adult females, leading to them recommending that the shear be less than 0.3 m/s, in order to produce maximum yield. Pace et al. (1986) used a kidney homogenate-yeast extract medium in which they inoculated X. nematophila 24 h before the inoculation of 2000 IJs/ml of S. carpocapsae.

Liquid culture technology was first made commercially available by the company Biosys, Palo Alto, California in 1992. The nematode produced was S. carpocapsae, which was upscaled to an 80 000-l fermenter. Currently, the majority of nematodes are produced in liquid culture by only a few companies, such as e-nema Gmbh (www.e-nema.de) in Germany, Koppert B.V. (www.koppert.nl) in The Netherlands, and Becker Underwood (www.beckerunderwood.com) and Certis (www.certisusa.com) from the United States (Ehlers & Shapiro-Ilan, 2005).

(32)

Liquid culture process technology

The long process time required for nematode cultures, combined with the need for an even distribution of fluids and organisms, result in the cultures being extremely vulnerable to contamination. Any non-symbiotic micro-organism that is present in the culture will lead to a reduction in nematode yield. Monoxenicity of nematode and bacterial cultures must be ensured from the outset in inoculum production.

Although symbiotic bacteria can be isolated from insect larvae infected with nematodes (Boemare & Akhurst, 1988), what is more laborious, however, is the establishing of bacteria-free nematodes. IJs cannot only be surface-sterilised as such a procedure does not exclude all contaminants. The preferred method requires the establishment of a monoxenic culture, by obtaining nematode eggs from gravid female stages (Lunau et al., 1993; Han & Ehlers, 1998). Monoxenic cultures of the nematode and bacteria together can be stored for months until they are inoculated into the bioreactor by shaking them at 20 rpm at 4°C.

Photorhabdus and Xenorhabdus have the ability to metabolise almost any kind of protein-rich medium. For this reason, selecting a specific medium for the nematode culture will depend mostly on economic considerations. A typical medium should have a carbon source (e.g. glucose or glycerol), various proteins of animal and plant origin, yeast extract, and, lastly, lipids of animal or plant origin (Pace et al., 1986; Friedman et al., 1989; Han et al., 1995; Surrey & Davies, 1996; Ehlers et al., 1998). Possible ways of increasing nematode yield entail either improving or adapting the liquid medium used (Ehlers, 2001).

Medium requirements vary between different nematode species. For example, S. carpocapsae require proteins of animal origin, without which they cannot reproduce (Yang et al., 1997). Production of offspring in a liquid medium without the addition of lipids is possible for H. bacteriophora, however (Han & Ehlers, 2001), because P. luminescens provides and metabolises the necessary lipids. S. glaseri is the only nematode for which essential amino acid requirements have been defined (Jackson, 1973). Nematodes can metabolise sterols from numerous steroid sources (Ritter, 1988), such as from lipids of either animal or plant origin. As a rule, though, lipids should always be added in order to increase the total

(33)

IJ fat content. The lipid composition affects the fatty acid composition of the bacteria and IJ (Abu Hatab et al., 1998), with a reduction in efficacy as a result of the low fat content in IJs (Patel et al., 1997a, 1997b).

Conventional equipment used in biotechnology, such as bioreactors with flat-blade impellers, bubble columns, and airlift, and internal-loop bioreactors, have been tested, with the latter consistently yielding higher IJ concentrations than have the other types (Ehlers & Shapiro-Ilan, 2005). Before the IJs are added, the medium is always inoculated and pre-incubated for 24–36 h with the nematode species specific symbiotic bacterium, which is normally between 0.5% and 1% of the total culture volume. The nematode inoculum usually forms between 5% and 10% of the total culture volume (Ehlers & Shapiro-Ilan, 2005).

Specific optimum culture temperature varies, depending on the nematode species and medium composition being used. The optimum temperature for the growth of a bacterial symbiont should always be defined before mass culturing of the nematode is attempted, as deviation from the optimum temperature can potentially induce a switch to the secondary phase, impeding nematode reproduction (Ehlers et al., 2000). When the culture medium is started, the pH should ideally be between 5.5 and 7, with the oxygen saturation rate being kept above 30%, as doing so will prevent the bacteria from switching to the secondary phase (Ehlers & Shapiro-Ilan, 2005).

One of the important parameters for an in vitro liquid culture is the aeration rate. Strauch and Ehlers (2000) compared yields of Heterorhabditis megidis Poinar, Jackson & Klein, 1987, with one culture being aerated at 0.3 vvm and another at 0.7 vvm. They found a significantly higher number of adults 8 days after IJ inoculation, and a higher final yield in the culture aerated at a higher rate. Increasing the aeration rate often leads to increased foaming, which silicon oil can be used to prevent, but high concentrations of such oil can affect the nematodes negatively. The use of long-chain fatty acids to control foaming was found to affect H. bacteriophora negatively (Ehlers & Shapiro-Ilan, 2005).

Numerous authors have reported final IJ yields from liquid culture (Pace et al., 1986; Bedding et al., 1993; Surrey & Davies, 1996; Han, 1996; Ehlers et al., 1998; Strauch & Ehlers, 2000). A negative correlation has been found to exist between the body length of the IJ and its yield. Body length is genetically defined and rather stable within a species, although it can differ according to strain and culture